Architecture

Application Flow Description

Data Collection

• Obtain vehicle driving behavior data (Sample data)
  • Development environment: Using simulated data
  • Production environment: Generated in real-time by sdv-flow and stored locally on the vehicle

Data Processing

1. Location Information Parsing

   • Extract latitude and longitude coordinates from driving behavior data
   • Parse corresponding administrative division information through Gaode MCP service

2. Weather Information Correlation

   • Based on administrative division information
   • Obtain real-time weather data at the time of the event through MCP Server which encapsulates third-party API services

Report Generation

• Integrate processed data (driving behavior, location, weather)
• Use large language models to analyze and generate UBI (Usage-Based Insurance) driving behavior reports

Running

Create Local .env File

# You could get the JuHe API key at https://www.juhe.cn/docs/api/id/277
JUHE_API_KEY=

# You could get the GaoDe API key at https://lbs.amap.com/api/mcp-server/create-project-and-key
GAODE_KEY=

# Silicon Flow API
SFAPI_KEY=

# Model Name
MODEL_NAME=Pro/deepseek-ai/DeepSeek-V3

# MQTT broker address
MQTT_BROKER=127.0.0.1

Initialize Environment

uv sync

source .venv/bin/activate

Execution

• Start emqx

docker run -d --name emqx -p 1883:1883 -p 8083:8083 -p 8084:8084 -p 8883:8883 -p 18083:18083 emqx/emqx:latest

• Run Weather MCP Server

uv run weather.py

• Run Vehicle Driving Behavior Analysis MCP Server

uv run vehicle.py

• Run Report Generation Script

uv run app.py

After executing app.py, a sample report will be generated.